立足于城市交通环卫的气象服务需求,利用近 65 年逐日气温和地温资料,结合智能网格总云量、气温数值预报和 Micaps地面天气形势资料及近两年逐日智能气象灾害监测站采集的路面温度数据和国家气象观测站日照时数(天空状况)和降水量实况数据,利用气候统计方法、天气学原理和 BP 神经网络算法,逐"候"精细统计分析了不同天空状况(晴天、多云、阴雨)、不同天气形势下气温与地面温度差值的变化特征,建立了 72 候气温与地面温度的差值定量指标库、智能网格地面温度预报和基于地面温度的路面温度预报模型,确定冬、夏半年五级路面状况预警等级指标,利用 matlab程序语言搭建基于气象条件的路面状况影响预报服务系统,为交通环卫开展路面温度数值预报和路面状况预警等级预报的专业气象服务.
Research on the Influence Prediction of Pavement Condition Based on Meteorological Conditions
To meet the meteorological service requirements of urban transportation and environmental sanitation,based on the daily temperature and ground temperature data in recent 65 years,the data of to-tal cloud cover and temperature numerical forecast of intelligent grid and the Micaps surface weather situ-ation data,as well as the daily road surface temperature data collected by intelligent meteorological disas-ter monitoring stations in recent two years,and the sunshine duration(sky condition)and precipitation data of national meteorological observation stations,we established the quantitative index database of difference between air temperature and ground temperature of 72 pentads,the intelligent grid ground tem-perature forecast model and the ground-temperature-based pavement temperature forecast model.Moreo-ver,by using climate statistical method,synoptic principle and BP neural network algorithm,we ana-lyzed the characteristics of difference between ground temperature and air temperature under different sky conditions(sunny,cloudy,rainy)and different weather conditions pentad by pentad.The index of five-level pavement condition warning grade in winter half year and summer half year was determined.The pavement condition influence prediction service system based on meteorological conditions was built by the matlab program language.These achievements can support the professional meteorological services of numerical pavement temperature prediction and pavement condition warning grade prediction for traffic and environmental sanitation.
meteorological conditionpavement conditionimpact predictionprediction test
万方数据
维普
